Molecular Dynamics Dataset for "Deep Learning-based Modeling Enhances Efficacy of Natural Ligand CAR Binders Targeting CD70"

CD27-CD70 N88A molecular dynamics and analysis The repository contains the following folders: 1. trajectories This folder contains all-atom molecular dynamics simulations of the trimeric CD27-CD70 complex and its N88A mutant. Each complex is simulated for 2.25 microseconds with three replicates. The directory holds the topologies and trajectories, with three replicates for each complex, and every analysis reads its input from here. The folders carry their own README that describes the files. 2. free_energy This folder builds the free energy landscapes of the two complexes, along with the root mean square deviation and the radius of gyration of the CA atoms. The script fes_2d.py writes fes_2d.png, and the script fes_3d.py writes fes_3d.png. 3. correlation_analysis This folder computes the linear mutual information and mutual information residue correlation networks with the MDIgest package. The pipeline runs 1_single_interface_analysis.ipynb and then 2_single_interface_plotting.ipynb. The notebook correlation_analysis.ipynb is a standalone robustness check that repeats the correlations across all three replicates. 4. interaction_analysis This folder counts hydrogen bonds, salt bridges, hydrophobic contacts, and general contacts at the mutation sites. The pipeline runs interaction_analysis.ipynb followed by interaction_plotting.ipynb, and the second notebook writes the figures and the manuscript tables. Running the analyses All analyses use three replicates, numbered I, II, and III. Each analysis folder contains its own README listing the inputs, run order, and outputs, and every analysis reads the trajectories from ../trajectories. Running these analyses requires the MDIgest package together with MDAnalysis, mdtraj, NumPy, SciPy, pandas, NetworkX, Matplotlib, and seaborn.